The aim of this study was to develop CPR by utilizing factors obtained from initial rehabilitation interventions that predict discharge outcomes in patients with cerebrovascular diseases during the acute phase.
This study was a retrospective observational study involving 119 patients who were hospitalized between January 1, 2023, and July 31, 2024, and diagnosed with acute stroke.
To predict discharge outcomes, eight candidate predictors were preselected and retrieved from electronic medical records: FIM cognitive subtotal, length of hospitalization, employment status, place of residence, FIM motor subtotal, sex, presence of a family member, and age.
Logistic regression analysis using the stepwise method was performed to reduce the number of predictors, requiring a significance level of 0.10 to enter the equation.
Receiver Operating Characteristic (ROC) curve analysis was used to calculate the cutoff values of the extracted independent variables for the outcomes.
For each variable, a score of 1 was assigned if the value was worse than the cutoff, and the CPR was created by summing these scores to evaluate their diagnostic characteristics.
Logistic regression analysis showed that the presence of a family member (odds ratio: 2.930, 95% CI: 0.943–9.130), FIM motor items (odds ratio: 0.957, 95% CI: 0.930–0.984), and FIM cognitive items (odds ratio: 0.944, 95% CI: 0.886–1.010) were identified as independent factors. Based on the ROC results, the cut-off value for the FIM motor items was 17 points (sensitivity: 0.6988, specificity: 0.9444), while for the FIM cognitive items, it was 27 points (sensitivity: 0.8434, specificity: 0.75).
The CPR consisted of the presence of a family member (yes = 0 points, no = 1 point), FIM motor items (≤17 points = 1 point, >17 points = 0 points), and FIM cognitive items (≤27 points = 1 point, >27 points = 0 points).
The most effective rule for predicting success was the presence of 2 or more of the 3 variables, indicating a 95.5% probability of a transfer to the hospital.
This study developed CPR to predict discharge outcomes based on FIM motor items, cognitive items, and the presence of a family member.
The importance of this study is that using this CPR allows for the prediction of discharge outcomes based on FIM assessment and patient background. Early prediction of discharge outcomes enables smoother adjustments for transfers and shorter lengths of stay in acute care hospitals.
Clinical prediction rule
discharge outcomes
